ABSTRACT The overarching aim of this grant proposal is to define and understand the conditions that create the best chance for cure or remission in HIV infection. There is a strong rationale for believing that in utero (IU) infection may provide the most promising starting point, because of the potential to initiate antiretroviral therapy (ART) within hours of birth, and because the immunotolerant environment in early life may reduce the ability of HIV to establish large viral reservoirs early in the course of infection. Aim 1 focuses on mechanisms of minimising the size of the viral reservoir in paediatric infection. Aim 2 addresses the ability of the effector arm of the paediatric immune system to eradicate HIV-infected cells as part of shock-and-kill strategies. Finally, to tackle the issue of how to identify children who will and who will not be suitable for ART treatment interruption, Aim 3 seeks to identify predictors of outcome post-treatment interruption. To approach these aims, we will study the cohorts of HIV-infected children we have generated over the past two decades of collaborations in South Africa. This includes two cohorts of HIV-infected children in whom ART was initiated in the neonatal period, one a historical cohort from 2002-2005, the other a current cohort in whom ART is initiated within 48hrs of birth. In addition we have identified in South Africa ~300 ART-nave children aged >5yrs who have maintained normal health and normal-for-age CD4 counts (>750 cells/mm3), whom we have termed `paediatric non-progressors' or `PNP'. These PNP represent approximately 5-10% of HIV-infected children. Some of these ART-nave children we have followed throughout childhood (ie from 0- 10yrs of age) and beyond, and these children still attend our clinics with their mothers today. These cohorts of children and mother-child pairs together provide us with unique opportunities to address our study aims that are fundamental to understanding how to achieve the best chance of HIV cure or remission. In Aim 1, Mechanisms of minimising the size of the viral reservoir in paediatric infection, we will determine the impact on the size of viral reservoir of ART initiation at 1-2 weeks of age ? the current standard of care in South Africa ? versus initiation within the first 1-2 days of life. In this sub-aim, we will build upon the infrastructure already established through our Wellcome Trust funded study started in 2015 in South Africa designed to establish the feasibility of point-of-care testing to diagnose in utero infection and initiated ART within the first 48hrs of life. We will also examine the infuence of viral factors such as viral replicative capacity - shown to be important in adults - on reservoir size, taking advantage of the opportunity to study transmitted virus within hours of birth. Finally we will test the hypothesis that paediatric non- progressors (defined above), in whom we have shown low levels of HIV infection in the long-lived Tscm and Tcm CD4 T-cell subsets compared to the Tem compartment, will show rapid and substantial decline in size of viral reservoirs on ART, as a result of the cellular localisation of HIV infection. In Aim 2, Mechanisms of optimal effector function for reservoir eradication, we argue that HIV-specific cytotoxic T lymphocytes (CTL) are likely to play an important role in eliminating HIV-infected cells that comprise the viral reservoir; however, control of viraemia is highly unusual in ART-nave paediatric infection. Among ~300 ART-nave paediatric non-progressors, we have identified <10% who have either reached undetectable viral loads (<20 copies/ml), or stable viral loads of <1000 c/ml. In contrast to adult elite controllers, control of viraemia in paediatric infection arises only after several years of infection, it is usually transient, and it appears unrelated to expression of protective HLA molecule such as HLA- B*57/58:01/81:01. Finally, unlike adults, HIV-infected children typically can generate CTL responses against CTL escape variants: potentilly this is the ultimate solution to HIV Study of these viraemic controllers longitudinally, in some cases from infancy, and comparison with viraemic non-controller children, will provide critical insights into the mechanisms of viraemic control in HIV-infected children. In Aim 3, Biomarkers predicting outcome post-treatment interruption, we propose the hypothesis that factors contributing to low viral reservoirs in the key, long-lived T cell subsets (Tscm and Tcm) also mediate HIV non-progression. In this aim we will first seek to identify features distinguishing progression from non- progression in longitudinally tracked children. In the second sub-aim, we will test the ability of these markers to predict outcome in a historical treatment interruption cohort of infants in whom ART was initiated at 4 weeks' age and interrupted after 12 months. Post-treatment interruption, 40% of infants progressed rapidly to restart ART within a median of 0.23yrs; in contrast, another 40% initiated ART >5yrs later.